b-open-io/cost-tracking
skills/cost-tracking/SKILL.md
This skill should be used when tracking API spend, analyzing billing data, monitoring service costs, running budget reports, or optimizing agent spend across Anthropic, Vercel, Railway, and other platforms. Invoke when asked about "cost tracking", "billing analysis", "spend report", "API costs", "usage monitoring", "budget vs actual", "cost optimization", or "which agents cost the most".
$ install --global
skillsauthnpx skillsauth add b-open-io/prompts cost-trackingInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jul 15, 2026, 4:43 AM128.7s4 files scanned
SKILL.md
- name:
- cost-tracking
- version:
- 1.0.2
- description:
- >-
Cost Tracking
Unified cost tracking and budget analysis across the bOpen agent organization's service stack.
Supported Platforms
Anthropic (Full API Access)
Usage API: GET /v1/organizations/usage_report/messages
- Auth: Admin API key (
sk-ant-admin...) — org admin required - Time buckets:
1m,1h,1d - Group by: model, workspace, api_key, service_tier, context_window_type, geo
- Data freshness: ~5 min lag
Cost API: GET /v1/organizations/cost_report
- Daily granularity only (
1d), costs in USD as decimal strings - Groups by workspace and description
- Includes token costs, web search, code execution
- Note: Priority Tier costs NOT included — use usage endpoint instead
Quick examples:
# Daily costs for last 7 days
curl -H "x-api-key: $ANTHROPIC_ADMIN_KEY" \
"https://api.anthropic.com/v1/organizations/cost_report?start_date=2026-03-07&end_date=2026-03-14&bucket_size=1d"
# Usage by model for last 24h
curl -H "x-api-key: $ANTHROPIC_ADMIN_KEY" \
"https://api.anthropic.com/v1/organizations/usage_report/messages?start_date=2026-03-13T00:00:00Z&end_date=2026-03-14T00:00:00Z&bucket_size=1h&group_by=model"
Full parameter reference and response shapes: references/anthropic-api.md.
Vercel (Full API Access)
Billing API: GET /v1/billing/charges
- Auth: Bearer token (same Vercel API token)
- Params:
from,to,teamIdorslug - Response: newline-delimited JSONL in the FOCUS version declared by the live Vercel API documentation; verify its field schema before parsing.
- Granularity: 1-day, max 1 year range
Spend Management Webhooks: fire at 50%/75%/100% of configured budget.
CLI shortcut: vercel usage for current-period spend at a glance.
Full API reference, FOCUS field descriptions, and webhook setup: references/vercel-api.md.
Railway (No API)
Railway has no public billing API. The GraphQL API covers projects, services, and deployments — not costs. Cost data is only available in the Railway dashboard under Settings > Usage.
When asked about Railway costs: instruct the user to open the dashboard and report the numbers manually. Do not attempt to fetch them programmatically.
Vantage MCP (Unified Dashboard)
Optional consolidated view across Anthropic, Vercel, and 30+ other providers.
claude mcp add --transport sse vantage https://mcp.vantage.sh/sse
Connects to Anthropic via Admin API key and to Vercel via a Community Integration with read:billing scope. Railway is NOT supported by Vantage.
Full setup steps: references/vantage-setup.md.
Cost Analysis Patterns
Per-Agent Cost Analysis
Query the Anthropic usage API grouped by api_key or workspace. Correlate workspace/key names with agent roles. Identify which agents burn the most tokens and recommend model tier changes (opus → sonnet → haiku) where quality allows.
Budget vs Actual
Pull cost data for the billing period from the Anthropic cost report and Vercel billing charges. Compare against budgeted amounts. Flag overruns with percentage and absolute delta.
Cost Anomaly Detection
Compare daily spend against the 7-day rolling average. Flag days where spend exceeds 2x the average. Drill into which model, workspace, or agent caused the spike by querying the usage endpoint grouped by model for that day.
Optimization Recommendations
- Identify agents using
claude-opusfor tasks thatclaude-sonnetorclaude-haikucould handle - Flag unused Vercel projects with non-zero spend
- Recommend caching strategies for repeated API calls
- Identify high-volume low-complexity calls suitable for Haiku downgrade
Key Principles
- Railway is manual-only. Never attempt programmatic Railway cost fetching — the API does not expose it.
- Anthropic cost report is daily-only. For hourly or per-minute analysis, use the usage endpoint, not the cost report.
- Priority Tier spend is missing from cost_report. Always note this limitation when reporting Anthropic totals.
- Vantage MCP is optional infrastructure. Useful for teams that want a persistent dashboard; not required for one-off queries.
References
references/anthropic-api.md— Full Anthropic Admin API docs for usage and cost endpointsreferences/vercel-api.md— Vercel Billing API, FOCUS format, and spend management webhooksreferences/vantage-setup.md— Vantage MCP server setup and platform coverage
Related Skills
b-open-io/claudex
tools
VerifiedTrustedCommunity
This skill should be used when a Claude Code session needs to keep working after Anthropic usage runs out, or when the user asks to run the Claude Code harness on GPT-5.6 Sol. Trigger phrases include "my Anthropic usage ran out", "I'm out of Claude usage", "usage limit reached, what now", "keep working on another model", "run Claude Code on GPT-5.6 Sol", "use GPT-5.6 Sol as the model", "set up claudex", "claudex isn't working", "route the harness through CLIProxyAPI", or "bill against my ChatGPT/Codex subscription". It stands up a local proxy so the Claude Code CLI runs on OpenAI's Codex backend as an escape hatch, and diagnoses that setup when it drifts. macOS + Homebrew.
14SKILL.mdUpdated Jul 17, 2026
b-open-io/visual-wayfinder
testing
VerifiedTrustedCommunity
This skill should be used when the user asks to "open Visual Wayfinder", "answer a Wayfinder ticket visually", "turn this decision into a configurator", "show Wayfinder choices as a dashboard", "prototype the Wayfinder questionnaire", or wants interactive choice cards, tradeoff controls, rankings, ranges, toggles, and consequence previews for one active Wayfinder decision. It wraps the Wayfinder skill and JSON Render; it never replaces the tracker or resolves more than the active decision.
14SKILL.mdUpdated Jul 16, 2026
b-open-io/visual-proposal
development
VerifiedTrustedCommunity
This skill should be used when the user asks to "make a visual proposal", "write this up so I can share it", "present these options visually", "diagram the trade-offs", "turn this plan into something reviewable", or requests a shareable design pitch, architecture proposal, RFC, options comparison, or visual roadmap for work that has not been built. It produces one self-contained, theme-aware HTML page led by grounded diagrams. Use visual-review instead for completed code changes; do not use this skill for internal task tracking.
14SKILL.mdUpdated Jul 16, 2026
b-open-io/plugin-settings
tools
VerifiedTrustedCommunity
This skill should be used when the user asks to "add plugin settings", "make a plugin configurable", "store per-project plugin configuration", "use settings.local.json", "create a plugin state file", "expose skill settings in Agent Master", or "add a skill interface". Distinguishes official Claude Code settings from project-owned configuration and documents bOpen Agent Master skill interface discovery.
14SKILL.mdUpdated Jul 16, 2026